FUJITA Toshiharu





1-1 Sensui-cho, Tobata-ku, Kitakyushu-shi, Fukuoka

Research Fields, Keywords

Dynamic programming


E-mail address

Undergraduate Education 【 display / non-display

  • 1992.03   Kyushu University   Faculty of Science   Graduated   JAPAN

Post Graduate Education 【 display / non-display

  • 1997.03  Kyushu University    Doctoral Program  Completed  JAPAN

Degree 【 display / non-display

  • Kyushu University -  Doctor of Mathematics  1997.03

Biography in Kyutech 【 display / non-display

  • 2017.02

    Kyushu Institute of TechnologyFaculty of Engineering   Department of Basic Sciences   Professor  

  • 2009.05

    Kyushu Institute of TechnologyScience Education Center   Associate Professor  

Specialized Field (scientific research fund) 【 display / non-display

  • Foundations of mathematics/Applied mathematics


Research Career 【 display / non-display

  • Nonserial Dynamic Programming Models

    dynamic programming, decision process, nonserial dynamics  

    Project Year: 2015.04  -  Now 

  • Mutually Dependent Decision Processes and its applications

    dynamic programming, decision process, nonserial dynamics  

    Project Year: 2010.04  -  Now 

  • Optimization with the Golden Number

    Golden Number, Optimization  

    Project Year: 2013.04  -  Now 

Publications (Article) 【 display / non-display

  • On Complementary Duals—Both Fixed Points—

    Iwamoto S., Kimura Y., Fujita T.

    Bulletin of the Kyushu Institute of Technology - Pure and Applied Mathematics    2020 ( 67 ) 1 - 28   2020.01  [Refereed]

     View Summary

    © 2020, Kyushu Institute of Technology. All rights reserved. We consider a quadratic minimization (primal) problem with both fixed endpoints and its associated maximization (dual) problem from a view point of complementarity. We focus on a pair of linear terms, which generates the respective quadratic functions (sums of squares) through an elementary inequality. We show that a complementary identity plays a fundamental part in establishing a dual relation between primal and dual. The identity produces the pair with an equality condition. The condition turns out to be a linear system of 2n-equation in 2n-variable. The system yields a couple of solutions, one is a minimum solution and the other is a maximum one. In the n-variable pair, both the solutions turn out to be complementary. The optimal solution is characterized by the backward Fibonacci sequence. The duality is enhanced through conjugate function. The solution is also given by dynamic programming. Thus Fibonacci complementary duality is established through the complementary identity approach.


  • 折り紙ユニットによる凸多面体構成問題について --- 実現可能な展開図の列挙 ---

    藤田 敏治

    京都大学数理解析研究所講究録    2126   106 - 115   2019.08  [Invited]

  • A dynamic programming algorithm for optimizing baseball strategies

    Kira Akifumi, Inakawa Keisuke, Fujita Toshiharu

    Journal of the Operations Research Society of Japan    62 ( 2 ) 64 - 82   2019.04  [Refereed]

     View Summary

    <p>In this paper, baseball is formulated as a finite non-zero-sum Markov game with approximately 6.45 million states. We give an effective dynamic programming algorithm which computes equilibrium strategies and the equilibrium winning percentages for both teams in less than 2 second per game. Optimal decision making can be found depending on the situation—for example, for the batting team, whether batting for a hit, stealing a base or sacrifice bunting will maximize their win percentage, or for the fielding team, whether to pitch to or intentionally walk a batter, yields optimal results. Based on this model, we discuss whether the last-batting team has an advantage. In addition, we compute the optimal batting order, in consideration of the decision making in a game.</p>

    repository DOI CiNii

  • 合流型推移をもつ決定過程について

    藤田 敏治, 才川 尚輝

    京都大学数理解析研究所講究録    2078   250 - 256   2018.07

  • Three recursive approaches for decision processes with a converging branch system

    Fujita T.

    Bulletin of the Kyushu Institute of Technology - Pure and Applied Mathematics    2018 ( 65 ) 1 - 21   2018.03

     View Summary

    © 2018, Kyushu Institute of Technology. All rights reserved. In this paper, we consider a decision process model with a converging branch system that is a nonserial transition system. The model is treated by three approaches. Thus we introduce three types of recursive equations by using a dynamic programming technique.


display all >>

Conference Prsentations (Oral, Poster) 【 display / non-display

  • 確率的推移をもつノンシリアル動的計画について


    第15回DP研究会  (長崎市)  2020.02  -  2020.02 

  • 微分と積分の話

    藤田敏治  [Invited]

    最適化法とその応用 第11回研究集会  (弘前)  2019.12  -  2019.12 

  • ノンシリアル動的計画について --- Feedforward Loop Systems ---

    藤田敏治  [Invited]

    最適化法とその応用 第11回研究集会  (弘前)  2019.12  -  2019.12 

  • 合流型の確率的推移をもつ決定過程問題について

    藤田敏治  [Invited]

    京都大学数理解析研究所研究集会:不確実・不確定性の下における数理的意思決定の理論と応用  (京都大学)  2019.11  -  2019.11 

  • 合流型推移をもつマルコフ決定過程


    本数学会2019年度年会  (東京工業大学)  2019.03  -  2019.03  日本数学会

display all >>